Worrawat Chanpattana, M.D.**
Wanchai Buppanharun, M.D.***
M.L. Somchai Chakrabhand, M.D****

Objective Determination of seizure threshold can help guide selection of stimulus dosage in electroconvulsive therapy (ECT); nonetheless, this threshold is subject to a variety of influences. This study aimed to measured the effect of the selection of ECT instrument on initial seizure threshold.

Method The initial seizure threshold was measured by stimulus dose titration in 88 patients, according to titration schedules with uniform increments. Treatment was given with the MECTA SR1 or the Thymatron DGx instrument, by random assignment, in groups with three age-related stratifications.

Results Measured seizure thresholds were higher with the stimuli used from the MECTA instrument than from the Thymatron instrument in 79% of patients (overall p < 0.0001), and were on average 61% higher (overall p < 0.0001).

instruments, dose-titration method, age and half-age methods

* Present at the 153^rd Annual Meeting of the American Psychiatric Association, McCormick Place Lakeside, Chicago, IL, USA, May 16, 2000.

*** Department of Preventive Medicine, Srinakharinwirot University, 681 Samsen, Dusit, Bangkok 10300.

**** Department of Mental Health, Ministry of Public Health, Tiwanont Road, Nonthaburi 11000.

วัตถุประสงค์ การหาปริมาณไฟฟ้าต่ำสุดที่ใช้ในการรักษาด้วยไฟฟ้ามีประโยชน์ในการเลือกใช้ปริมาณไฟฟ้าที่เหมาะสมในการรักษา งานวิจัยนี้ทำการศึกษาเปรียบเทียบผลของเครื่องมือที่แตกต่างกันที่มีต่อปริมาณไฟฟ้าต่ำสุดที่ใช้เริ่มต้นการรักษาในผู้ป่วยจิตเภทและผู้ป่วย schizoaffective

วิธีการศึกษา ศึกษาผู้ป่วย schizoaffective จำนวน 88 คน โดยใช้เกณฑ์ปรับปริมาณไฟฟ้าของมหาวิทยาลัยศรีนครินทรวิโรฒ ผู้ป่วยถูกแบ่งออกเป็น 3 กลุ่มอายุ และสุ่มให้ได้รับการรักษาด้วยเครื่องมือ 2 ชนิด ได้แก่ MECTA SR1 และ Thymatron DGx

ผลการศึกษา ค่าปริมาณไฟฟ้าต่ำสุดที่วัดได้จาก MECTA SR1 มากกว่าค่าที่ได้จาก Thymatron DGx ถึงร้อยละ 61 ในผู้ป่วยร้อยละ 79

สรุป เนื่องจากการใช้ปริมาณไฟฟ้าสูงทำให้เกิดผลข้างเคียงต่อความจำมากขึ้น ความแตกต่างนี้อาจส่งผลต่อการรักษาด้วยไฟฟ้าในเวชปฏิบัติ และอาจต้องพิจารณาปรับปริมาณไฟฟ้าให้เหมาะสมเมื่อมีการเปลี่ยนเครื่องมือที่ใช้ในระหว่างการรักษา

วารสารสมาคมจิตแพทย์แห่งประเทศไทย 2543; 45(2): 145-153.

คำสำคัญ การรักษาด้วยไฟฟ้า ปริมาณไฟฟ้าต่ำสุดที่ใช้ในการรักษา ผลของเครื่องมือที่ใช้รักษาด้วยไฟฟ้า

* ภาควิชาจิตเวชศาสตร์ คณะแพทยศาสตร์ มหาวิทยาลัยศรีนครินทรวิโรฒ ถนนสามเสน ดุสิต กรุงเทพฯ 10300

** ภาควิชาเวชศาสตร์ป้องกัน คณะแพทยศาสตร์ มหาวิทยาลัยศรีนครินทรวิโรฒ ถนนสามเสน ดุสิต กรุงเทพฯ 10300

*** กรมสุขภาพจิต กระทรวงสาธารณสุข ถนนติวานนท์ นนทบุรี 11000

Estimation of the seizure threshold can help guide the selection of the electrical stimulus dose at electroconvulsive therapy (ECT)¹. In concept, this threshold is the smallest dose of electrical charge that can induce a seizure². In practice this minimum dose depends not only on individual patient characteristics and treatment method, but also on several aspects of stimulus waveform. Examples of the former include electrode placement³, anesthetic agents¹ and concomitant medications⁴, age and sex⁵, and the frequency of ECT sessions⁶. Stimulus waveform aspects that influence seizure threshold include pulsewidth, charge rate^7,8, and waveshape such as square or sine^3,9; differences in these between modern ECT instruments are the focus of the present study.

The accurate communication of clinical issues and research results in ECT requires an understanding of how they apply with other common instrumentation. The only commercially available ECT instruments to incorporate the EEG monitoring features recommended by the APA Task Force on ECT (2000) are the MECTA and Somatics Thymatron instruments¹, and virtually all modern ECT publications note the use of one of these. The instrument models we used are usually cited in recent studies, and so are representative. Although they both deliver constant current brief-pulse stimuli, the ranges of stimulus parameters and the method of stimulus selection differ. A fundamental issue is the nature of the correspondence between the instruments about stimulus settings, particularly the charge dose. An associated concern is how much the instruments differ in the efficiency of the stimuli they deliver.

We could find no published study that compared seizure thresholds with different ECT instruments. We conducted a prospective, randomized controlled trial to compare initial seizure threshold estimated by stimulus dose-titration technique with the MECTA SR1 and Thymatron DGx instruments.

The subjects were 88 patients hospitalized for acute exacerbation of psychosis who were selected to receive ECT on clinical grounds at the participating hospitals. Each met DSM-IV¹⁰ criteria for schizophrenia (n = 75) or schizoaffective disorder (n = 13). The study was approved by the Ethics Committee of the Faculty of Medicine of Srinakharinwirot University. After a detailed explanation, subjects and/or their guardians gave written informed consent for ECT and for study participation. We excluded subjects who had received ECT or depot neuroleptics within six months or were taking medicines that inhibit seizure, e.g., anticonvulsants, benzodiazepines, beta-blockers.

From the outset subjects were stratified by age into three groups: 30 or less, 31 to 40, and over 40 years. They were randomly assigned to receive ECT with either the MECTA SR1 or the Thymatron DGx instrument. All subjects were free of medicines beginning 5 days prior to the first ECT, except for flupenthixol 12 mg/day and benzhexol 4-6 mg/day, which all received. All data were collected during the first two ECT sessions, which were given two to three days apart.

For both MECTA and Thymatron groups, average subject age was 38.2 years, with 9.6 years of education; 14 males and 30 females were in each group. Differences between groups were negligible for onset of illness (average 19.9 + 3.4 years), illness duration (17.8 + 9.5 years), episode duration (1.6 + 1.5 years), numbers of admissions (8.6 + 4.8 years), percent with prior ECT (85%), entry BPRS score (48.2 + 8.9), entry GAF score (31.6 + 6.3), and percent with schizoaffective disorder (15%).

After atropine 0.4 mg intravenously, anesthesia was given with a minimal dosage of thiopental (2-4 mg/kg) and 0.5-1 mg/kg of succinylcholine. Subjects were hyperventilated with oxygen from anesthetization until postictal spontaneous respiration.

Bitemporal bilateral electrode placement was used exclusively. Motor seizure activity was monitored by the cuffed ankle method⁴, and two channels of prefrontal electroencephalogram (EEG) were recorded from frontal and mastoid electrodes.

Seizure threshold was measured according to a titration schedule (Table 1) at the first and second treatment sessions. This schedule incorporated the Thymatron factory default settings, as representative of it. The MECTA has no default or standard settings specified by its manufacturer; its settings were chosen to match the method of the Thymatron. In this the stimulus charge is expressed as its percentage of the instrument’s maximum, in equal increments of 5% from 5% to 100%, and is referred to as “%Energy.” In diminishing priority order we then matched current, pulsewidth, and frequency. Uniform increments of stimulus dose contribute to the systematic and impartial measurement of seizure threshold. This dose method is the only one we know with reasonably uniform increments for the MECTA SR1; Thymatron dose settings have uniform increments. Because these stimuli are nonproprietary, and devices can change, our usage of the terms “Thymatron” and “MECTA” refers to the configurations of stimulus parameters we studied rather than inevitably these instruments.

Operationally for study purposes we defined an adequate seizure as bilateral tonic-clonic motor activity that lasted for at least 30 seconds, together with EEG evidence of seizure. Accordingly, the thresholds we measured are for vigorous rather than minimal seizures¹¹. At the first treatment session, the first level of stimulus intensity (10% of maximum charge) was administered. If this failed to elicit an adequate seizure the stimulus charge was increased in increments of 10% Energy as listed in Table 1. A maximum of four stimulations per session was allowed, with an interval of at least 20 seconds (for missed seizure) or 40 seconds (short seizure) between stimulations. Additional thiopental was not administered. At the second treatment session for each subject, stimulus dose lower by 5% Energy than at the first session was given, as listed in Table 1. If an adequate seizure occurred, that dose was taken as the threshold; if not, the first session’s stimulus dose was so taken.

Seizure threshold expressed in millicoulombs (mC) were transformed logarithmically to increase the normality of the distribution. Separately, seizure threshold was analyzed in %Energy units. Differences between groups on single continuous variables were evaluated by t test or analysis of variance (ANOVA). Relationships between continuous variables were characterized by Pearson’s product-moment correlation coefficient. Prediction of seizure threshold was examined by stepwise multiple regression analysis. Values are given as mean + SD. SPSS (1996 SPSS Inc.) was used.

The seizure threshold varied from 25.2 to 252 mC, with an overall mean of 103.1 + 45.5 mC. There was a nonsignificant trend for higher threshold in women than men [108.8 + 46 mC, n = 60, vs. 90.9 + 42.7 mC, n = 28; t (86) = 1.85, p = 0.068], consistent with their older age [40.7 + 10.0 vs. 32.8 + 8.9 years; t (86) = 3.59, p = 0.001].

Seizure threshold charge was on average 61% higher with the MECTA SR1 than the Thymatron DGx over the entire sample; the threshold was significantly higher in each age group, as shown in Figure 1. There were no significant differences in motor seizure duration (49.7 + 14.1s MECTA vs. 52.1 + 15.1s Thymatron, t = 0.77, p = 0.45), EEG seizure duration (63.9 + 34.2s vs. 62.4 + 19.2s, t = 0.25, p = 0.81), or in doses of thiopental (141.5 + 24.7 mg vs. 144.3 + 30 mg, t = 0.49, p = 0.63) or succinylcholine (27.3 + 14 mg vs. 24.5 + 6.2 mg, t = 1.23, p = 0.22).

Seizure threshold correlated with age (r = 0.51, p < 0.0001), illness duration (r = 0.54, p < 0.0001) and ECT instrument (Spearman’s r = 0.46, p < 0.0001; Thymatron = 1, MECTA = 2). Stepwise multiple regression analysis, using probability-of-F < 0.05 to enter and > 1.0 to delete, revealed that illness duration (t = 6.1, p < 0.0001) followed by instrument [t = 4.5, p < 0.0001; F (2,85) = 31.69, p < 0.0001] contributed to seizure threshold; these variables accounted for 42.7% of total variance.

Expression of the seizure threshold in ‘% Energy’ units produced similar results, with MECTA groups showing higher thresholds than Thymatron groups [overall: F (1,86) = 5.41, p = 0.022]. Note that at any specific ‘% Energy’ value, the charge with the MECTA instrument is about 15% higher than with the Thymatron instrument.

Because the observations of lower seizure threshold with the Thymatron DGx than the MECTA SR1 in three separate groups constitute three independent trials, and each trial produced statistical significance, the overall statistical significance is the product of the three separate results, which is p < 0.0001, F (1,86) = 18.38. This result is logically consistent with the 93% success rate of the half-age method in identifying a valid dose for Thymatron stimuli, compared to its 68% success rate for MECTA stimuli. It is also consistent with the smaller number of stimulations needed to induce a seizure with Thymatron stimuli. Our intention in making three stratifications by age was to determine if the results varied substantially by age; they do not, despite an approximate 40% effect of age on seizure threshold.

Underlying the results are a variety of differences in stimulus characteristics, and presumably the greater efficiency associated with stimuli of lower charge rate⁸, lower pulsewidth¹⁴, lower pulse frequency¹⁵, and longer train duration^7,15. The present study did not examine individual stimulus parameters, but rather compared sets of stimuli that represent uniform increment stimulus titration with different instruments. The highest stimulus that failed to induce seizure for each subject was tallied in Table 2. Differences in these highest-failure stimuli represent differences between instruments. As seen in this table, with the MECTA instrument larger numbers of subjects appear at doses of 20%, 25% and 35% Energy. The numbers of subject differences between instruments at these stimuli are 4, 6, and 2, respectively. The MECTA stimuli at these doses do not appear unusual or unrepresentative of this instrument.

The higher seizure threshold shown in our female patients probably follows from their older average age [40.7 + 10.0 vs. 32.8 + 8.9 years, t (86) = 3.59, p = 0.001]. Similarly, the relationship between seizure threshold and illness duration presumably follows the correlation between illness duration and age (r = 0.93, p < 0.0001).

The differences observed between the sets of ECT stimuli we compared have potential clinical implications, because of the reasonable expectation of associations between greater cognitive side effects and lower stimulus efficiency, i.e., higher seizure threshold⁸. As evidence of this, widephase sinusoidal stimuli (e.g., 8.3 ms phasewidth = 60 Hz) cause greater adverse cognitive effects¹⁶ and have a several-fold higher seizure threshold³ than brief-pulse stimuli. Sackeim et al. (1991) note that different stimuli of the same dose might differ in cognitive consequences¹⁷, but did not compare cognitive side effects in patients with different thresholds or receiving stimuli with different stimulus waveform characteristics.

Another clinical consideration is adjustment of the stimulus dose if there is a change in the ECT instrument used with a patient. The present results suggest that an increase of about 60% in the stimulus charge is likely needed to avoid failure of seizure induction when changing from the configuration of stimulus parameters represented by the Thymatron instrument to that represented by the MECTA instrument.

This study is supported in part by the Thailand Research Fund, grant BRG 3980009. We thank Wiwat Yatapootanon, M.D. for technical assistance.

1. American Psychiatric Association Task Force Report. The practice of ECT:

Recommendations for treatment, training, and privileging. Washington, DC: American Psychiatric Press, 2000 (in press).

2. Small JG, Small IF, Milstein V. Electrophysiology in ECT. In: Lipton MA, DiMascio A, Killam KF (eds.). Psychopharmacology: A generation of progress. New York: Raven Press, 1978: 759-69.

3. Weiner RD. ECT and seizure threshold: Effects of stimulus waveform and electrode

placement. Biol Psychiatry 1980; 15: 225-41.

4. Kellner CH, Pritchett JT, Beale MD, Coffey CE. Handbook of ECT. Washington, DC: American Psychiatric Press, 1997: 64.

5. Sackeim HA, Decina P, Prohovnik I, Malitz S. Seizure threshold in electroconvulsive

therapy: Effects of sex, age, electrode placement, and number of treatments. Arch Gen Psychiatry 1987; 44: 355-60.

6. Janakiramiah N, Jyotti RKM, Praveen J, et al. Seizure duration over ECT sessions:

Influence of spacing ECTs. Indian J Psychiatry 1992; 34: 124-7.

7. Swartz CM, Larson G. ECT stimulus duration and its efficacy. Ann Clin Psychiatry 1989; 1: 147-152.

8. Swartz CM. Optimizing the ECT stimulus. Convulsive Ther 1994; 10: 132-4.

9. Weaver LA, Ives JO, Williams R, et al. A comparison of standard alternating current and low-energy brief-pulse electrotherapy. Biol Psychiatry 1977; 12: 525-43.

10. American Psychiatric Association. Diagnostic and Statistical Manual of Mental

Disorders, 4^th ed. Washington, DC: American Psychiatric Press, 1994.

11. Christensen P, Kragh Sorensen P, Sorensen C, et al. EEG-monitored ECT: A comparison of seizure duration under anesthesia with etomidate and thiopentone. Convulsive Ther 1986; 2: 145-50.

12. Swartz CM, Abrams R. ECT instruction manual. Illinois: Somatics Inc., 1989.

13. Petrides G, Fink M. The “half-age” stimulation strategy for ECT dosing. Convulsive Ther 1996; 12: 138-46.

14. Swartz CM, Manly DT. ECT pulsewidth 0.5 millisecond is more efficient than 1.0

millisecond stimuli [abstract]. In Proceedings of the 149^th APA Annual Meeting, San Diego, 1997; New Research Abstract No. 237, p. 132.

15. Devanand DP, Lisanby SH, Nobler MS, et al. The relative efficiency of altering pulse

frequency or train train duration when determining seizure threshold. JECT 1998; 14: 227-35.

16. Weiner RD, Roger HJ, Davidson JRT, Squire LR. Effects of stimulus parameters on

cognitive side effects. Ann NY Acad Sci 1986; 462: 315-25.

17. Sackeim HA, Devanand DP, Prudic J. Stimulus intensity, seizure threshold, and

seizure duration: impact on the efficacy and safety of electroconvulsive therapy.

Psychiatr Clin North Am 1991; 14: 803-43.

Table 1. The stimuli used for titration

% PW Freq Duration I Charge rate Charge Charge Freq Duration Charge rate

(ms) (Hz) (s) (A) (mC/s) (mC) (mC) (Hz) (s) (mC/s)

5 1.0 40 0.5 0.8 64 32 25 30 0.47 54

10 1.0 40 1.25 0.6 48 60 50 30 0.93 54

15 1.0 40 1.5 0.7 56 84 76 30 1.4 54

20 1.0 40 2.0 0.75 60 120 101 30 1.87 54

25 1.0 90 1.0 0.8 144 144 126 30 2.33 54

30 1.0 60 2.0 0.75 90 180 151 50 1.68 90

35 1.0 60 2.0 0.8 96 192 176 50 1.96 90

40 1.2 60 2.0 0.8 115 230 202 50 2.24 90

45 1.2 70 2.0 0.75 126 252 227 50 2.52 90

50 1.0 90 2.0 0.8 144 288 252 50 2.8 90

Abbreviations: PW = pulsewidth; Freq = frequency; I = current

Table 2. Tally of highest stimuli that failed to induce seizure, per subject*.

* The seizure threshold was 5% higher. On MECTA 10% = 60 mC,

on Thymatron 10% = 50.4 mC.